lock0lock.h, lock0lock.c:

  Remove a spurious deadlock when >= 2 insert transactions have to wait because there is a next-key lock set


innobase/lock/lock0lock.c:
  Remove a spurious deadlock when >= 2 insert transactions have to wait because there is a next-key lock set
innobase/include/lock0lock.h:
  Remove a spurious deadlock when >= 2 insert transactions have to wait because there is a next-key lock set

innobase/include/lock0lock.h

@@ -537,7 +537,15 @@ extern lock_sys_t*	lock_sys;
 				the bit is set; locks of this type are created
 				when records are removed from the index chain
 				of records */
-
+#define LOCK_INSERT_INTENTION 1024 /* this bit is set when we place a waiting
+				gap type record lock request in order to let
+				an insert of an index record to wait until
+				there are no conflicting locks by other
+				transactions on the gap; note that this flag
+				remains set when the waiting lock is granted,
+				or if the lock is inherited to a neighboring
+				record */
+				
 /* When lock bits are reset, the following flags are available: */
 #define LOCK_RELEASE_WAIT	1
 #define LOCK_NOT_RELEASE_WAIT	2

innobase/lock/lock0lock.c

@@ -96,7 +96,9 @@ is only performed when a transaction itself requests a lock!
 
 An insert is allowed to a gap if there are no explicit lock requests by
 other transactions on the next record. It does not matter if these lock
-requests are granted or waiting, gap bit set or not. On the other hand, an
+requests are granted or waiting, gap bit set or not, with the exception
+that a gap type request set by another transaction to wait for
+its turn to do an insert is ignored. On the other hand, an
 implicit x-lock by another transaction does not prevent an insert, which
 allows for more concurrency when using an Oracle-style sequence number
 generator for the primary key with many transactions doing inserts
@@ -544,8 +546,7 @@ lock_sec_rec_cons_read_sees(
 				index record */
 	rec_t*		rec,	/* in: user record which should be read or
 				passed over by a read cursor */
-	dict_index_t*	index __attribute__((unused)), 
-                                /* in: non-clustered index */
+	dict_index_t*	index,	/* in: non-clustered index */
 	read_view_t*	view)	/* in: consistent read view */
 {
 	dulint	max_trx_id;
@@ -616,7 +617,7 @@ UNIV_INLINE
 ulint
 lock_get_type(
 /*==========*/
-			/* out: LOCK_TABLE or LOCK_RECa */
+			/* out: LOCK_TABLE or LOCK_REC */
 	lock_t*	lock)	/* in: lock */
 {
 	ut_ad(lock);
@@ -717,6 +718,46 @@ lock_rec_set_gap(
 	}
 }
 
+/*************************************************************************
+Gets the waiting insert flag of a record lock. */
+UNIV_INLINE
+ibool
+lock_rec_get_insert_intention(
+/*==========================*/
+			/* out: TRUE if gap flag set */
+	lock_t*	lock)	/* in: record lock */
+{
+	ut_ad(lock);
+	ut_ad(lock_get_type(lock) == LOCK_REC);
+
+	if (lock->type_mode & LOCK_INSERT_INTENTION) {
+
+		return(TRUE);
+	}
+
+	return(FALSE);
+}
+
+/*************************************************************************
+Sets the waiting insert flag of a record lock. */
+UNIV_INLINE
+void
+lock_rec_set_insert_intention(
+/*==========================*/
+	lock_t*	lock,	/* in: record lock */
+	ibool	val)	/* in: value to set: TRUE or FALSE */
+{
+	ut_ad(lock);
+	ut_ad((val == TRUE) || (val == FALSE));
+	ut_ad(lock_get_type(lock) == LOCK_REC);
+
+	if (val) {
+ 		lock->type_mode = lock->type_mode | LOCK_INSERT_INTENTION;
+	} else {
+		lock->type_mode = lock->type_mode & ~LOCK_INSERT_INTENTION;
+	}
+}
+
 /*************************************************************************
 Calculates if lock mode 1 is stronger or equal to lock mode 2. */
 UNIV_INLINE
@@ -1331,6 +1372,12 @@ lock_rec_other_has_expl_req(
 			into account, or 0 if not */
 	ulint	wait,	/* in: LOCK_WAIT if also waiting locks are
 			taken into account, or 0 if not */
+	ibool	ignore_insert_intention,/* in: FALSE if also locks where the
+			LOCK_INSERT_INTENTION
+			flag is set should be taken into account; we can
+			use this option to prevent unnecessary deadlocks
+			between inserts which have needed to set a wait
+			lock to wait for some next-key lock to be released */
 	rec_t*	rec,	/* in: record to look at */	
 	trx_t*	trx)	/* in: transaction, or NULL if requests
 			by any transaction are wanted */
@@ -1338,8 +1385,11 @@ lock_rec_other_has_expl_req(
 	lock_t*	lock;
 	
 	ut_ad(mutex_own(&kernel_mutex));
-	ut_ad((mode == LOCK_X) || (mode == LOCK_S));
-
+	ut_ad(mode == LOCK_X || mode == LOCK_S);
+	ut_ad(gap == 0 || gap == LOCK_GAP);
+	ut_ad(wait == 0 || wait == LOCK_WAIT);
+	ut_ad(ignore_insert_intention == TRUE
+					|| ignore_insert_intention == FALSE);
 	lock = lock_rec_get_first(rec);
 
 	while (lock) {
@@ -1349,7 +1399,14 @@ lock_rec_other_has_expl_req(
 		    && (wait || !lock_get_wait(lock))
 		    && lock_mode_stronger_or_eq(lock_get_mode(lock), mode)) {
 
-		    	return(lock);
+		    	if (ignore_insert_intention
+				&& lock_rec_get_insert_intention(lock)) {
+				
+				/* We can ignore this lock */
+			} else {
+
+		    		return(lock);
+		    	}
 		}
 
 		lock = lock_rec_get_next(rec, lock);
@@ -1526,7 +1583,10 @@ lock_rec_enqueue_waiting(
 				DB_QUE_THR_SUSPENDED */
 	ulint		type_mode,/* in: lock mode this transaction is
 				requesting: LOCK_S or LOCK_X, ORed with
-				LOCK_GAP if a gap lock is requested */
+				LOCK_GAP if a gap lock is requested, ORed
+				with LOCK_INSERT_INTENTION if the waiting
+				lock request is set when performing an
+				insert of an index record */
 	rec_t*		rec,	/* in: record */
 	dict_index_t*	index,	/* in: index of record */
 	que_thr_t*	thr)	/* in: query thread */
@@ -1613,11 +1673,12 @@ lock_rec_add_to_queue(
 	ut_ad(mutex_own(&kernel_mutex));
 	ut_ad((type_mode & (LOCK_WAIT | LOCK_GAP))
 	      || ((type_mode & LOCK_MODE_MASK) != LOCK_S)
-	      || !lock_rec_other_has_expl_req(LOCK_X, 0, LOCK_WAIT, rec, trx));
+	      || !lock_rec_other_has_expl_req(LOCK_X, 0, LOCK_WAIT,
+						FALSE, rec, trx));
 	ut_ad((type_mode & (LOCK_WAIT | LOCK_GAP))
 	      || ((type_mode & LOCK_MODE_MASK) != LOCK_X)
-	      || !lock_rec_other_has_expl_req(LOCK_S, 0, LOCK_WAIT, rec, trx));
-	      
+	      || !lock_rec_other_has_expl_req(LOCK_S, 0, LOCK_WAIT,
+						FALSE,rec, trx));
 	type_mode = type_mode | LOCK_REC;
 
 	page = buf_frame_align(rec);
@@ -1752,8 +1813,8 @@ lock_rec_lock_slow(
 		nothing */
 
 		err = DB_SUCCESS;
-	} else if (lock_rec_other_has_expl_req(confl_mode, 0, LOCK_WAIT, rec,
-								trx)) {
+	} else if (lock_rec_other_has_expl_req(confl_mode, 0, LOCK_WAIT,
+							FALSE, rec, trx)) {
 		/* If another transaction has a non-gap conflicting request in
 		the queue, as this transaction does not have a lock strong
 		enough already granted on the record, we have to wait. */
@@ -2056,7 +2117,8 @@ lock_rec_inherit_to_gap(
 	lock = lock_rec_get_first(rec);
 
 	while (lock != NULL) {
-		lock_rec_add_to_queue((lock->type_mode | LOCK_GAP) & ~LOCK_WAIT,
+		lock_rec_add_to_queue(((lock->type_mode | LOCK_GAP)
+					& ~LOCK_WAIT),
 	 			     		heir, lock->index, lock->trx);
 		lock = lock_rec_get_next(rec, lock);
 	}
@@ -2312,7 +2374,7 @@ list start is moved to another page. */
 void
 lock_move_rec_list_start(
 /*=====================*/
-	page_t*	new_page __attribute__((unused)),/*in: index page to move to */
+	page_t*	new_page,	/* in: index page to move to */
 	page_t*	page,		/* in: index page */
 	rec_t*	rec,		/* in: record on page: this is the
 				first record NOT copied */
@@ -2782,7 +2844,7 @@ lock_deadlock_recursive(
 				we return TRUE */
 {
 	lock_t*	lock;
-	ulint	bit_no = 0; /* remove warning */
+	ulint	bit_no;
 	trx_t*	lock_trx;
 	char*	err_buf;
 	
@@ -3486,6 +3548,10 @@ lock_rec_print(
 		buf += sprintf(buf, " gap type lock");
 	}
 
+	if (lock_rec_get_insert_intention(lock)) {
+		buf += sprintf(buf, " insert intention");
+	}
+
 	if (lock_get_wait(lock)) {
 		buf += sprintf(buf, " waiting");
 	}
@@ -3503,6 +3569,15 @@ lock_rec_print(
 					IB__FILE__, __LINE__, &mtr);
 	if (page) {
 		page = buf_page_get_nowait(space, page_no, RW_S_LATCH, &mtr);
+
+		if (!page) {
+			/* Let us try to get an X-latch. If the current thread
+			is holding an X-latch on the page, we cannot get an
+			S-latch. */
+			
+			page = buf_page_get_nowait(space, page_no, RW_X_LATCH,
+									&mtr);
+		}
 	}
 				
 	if (page) {
@@ -3869,7 +3944,7 @@ lock_rec_queue_validate(
 		impl_trx = lock_clust_rec_some_has_impl(rec, index);
 
 		if (impl_trx && lock_rec_other_has_expl_req(LOCK_S, 0,
-						LOCK_WAIT, rec, impl_trx)) {
+				LOCK_WAIT, FALSE, rec, impl_trx)) {
 
 			ut_a(lock_rec_has_expl(LOCK_X, rec, impl_trx));
 		}
@@ -3884,7 +3959,7 @@ lock_rec_queue_validate(
 		impl_trx = lock_sec_rec_some_has_impl_off_kernel(rec, index);
 
 		if (impl_trx && lock_rec_other_has_expl_req(LOCK_S, 0,
-						LOCK_WAIT, rec, impl_trx)) {
+				LOCK_WAIT, FALSE, rec, impl_trx)) {
 
 			ut_a(lock_rec_has_expl(LOCK_X, rec, impl_trx));
 		}
@@ -3909,12 +3984,10 @@ lock_rec_queue_validate(
 		
 			if (lock_get_mode(lock) == LOCK_S) {
 				ut_a(!lock_rec_other_has_expl_req(LOCK_X,
-								0, 0, rec,
-								lock->trx));
+						0, 0, FALSE, rec, lock->trx));
 			} else {
 				ut_a(!lock_rec_other_has_expl_req(LOCK_S,
-								0, 0, rec,
-								lock->trx));
+						0, 0, FALSE, rec, lock->trx));
 			}
 
 		} else if (lock_get_wait(lock) && !lock_rec_get_gap(lock)) {
@@ -4153,11 +4226,19 @@ lock_rec_insert_check_and_lock(
 	*inherit = TRUE;
 
 	/* If another transaction has an explicit lock request, gap or not,
-	waiting or granted, on the successor, the insert has to wait */
-
-	if (lock_rec_other_has_expl_req(LOCK_S, LOCK_GAP, LOCK_WAIT, next_rec,
-								trx)) {
-		err = lock_rec_enqueue_waiting(LOCK_X | LOCK_GAP, next_rec,
+	waiting or granted, on the successor, the insert has to wait.
+
+	An exception is the case where the lock by the another transaction
+	is a gap type lock which it placed to wait for its turn to insert. We
+	do not consider that kind of a lock conflicting with our insert. This
+	eliminates an unnecessary deadlock which resulted when 2 transactions
+	had to wait for their insert. Both had waiting gap type lock requests
+	on the successor, which produced an unnecessary deadlock. */
+
+	if (lock_rec_other_has_expl_req(LOCK_S, LOCK_GAP, LOCK_WAIT,
+						TRUE, next_rec, trx)) {
+		err = lock_rec_enqueue_waiting(LOCK_X | LOCK_GAP
+					| LOCK_INSERT_INTENTION, next_rec,
 								index, thr);
 	} else {
 		err = DB_SUCCESS;